Step 5 Preview scores 54.9 on the Respan Index (#12), ahead of DeepSeek V4.1 Flash at 50.9 (#20). DeepSeek V4.1 Flash leads on 3 of the 4 benchmarks both report. DeepSeek V4.1 Flash is 2.7x cheaper per token at list price.
| Benchmark | DeepSeek V4.1 Flash | Step 5 Preview |
|---|---|---|
| Terminal-Bench 4.0 (Vals AI) | 19.7% | 31.8% |
| LMArena Agent (LMArena) | 0.0402 | 0.0048 |
| LMArena Elo | 1474 | 1454 |
| LMArena WebDev (LMArena) | 1620 | 1570 |
Higher is better. Each score links to where it was published.
* Estimated: no published scores in that category. How the Respan Index works